Petroleum refiners have been removing salts from crude oil and product fractions almost since the dawn of refining.
Crude oil contains salts and salt precursors such as nitrogen and sulfur compounds. Salt is removed upstream of fractionation equipment to prevent salt deposition in columns and associated equipment. Simple water washing will remove these salts, and most refiners now use "desalters."
Modern refining techniques also produce salts or sometimes introduce salts into process streams.
The main source of the produced salts is chemical reactions between an acidic or basic compound dissolved in the oil and an acid or base added or created during refinery processing. Thus nitrogen and sulfur compounds in the feed are frequently converted to hydrogen sulfide or ammonia. Another impurity, hydrogen chloride, can be produced by hydrolysis of calcium and magnesium chlorides or by hydrogenation of organic chlorides. Ammonia will react with hydrogen sulfide and hydrogen chloride to produce ammonium hydrosulfide and/or ammonium chloride, respectively.
Salts can be introduced into a refinery stream when a catalyst contains leachable alkaline or acidic components. Thus platinum reforming catalyst can introduce chlorine into reformate and to the reformer off gas.
These salts, regardless of source, can form either solid deposits or an aqueous phase if water is present.
Although salt deposits, or formation of salt solutions, might seem like minor problems the opposite is the case. Salt deposits can plug distillation column internals, and shut down an entire refinery. The salts are also hygroscopic and attract water to form extremely corrosive regions. Corrosive salt solutions can and do eat holes in refinery vessels.
As an example of how serious the problem can be, on May 5, 1988 at about 3:30 a.m. an explosion occurred at a Louisiana refinery which resulted in 7 fatalities, 28 injuries, and significant property damage.
A depropanizer associated with a large FCC unit had experienced condenser fouling. To control fouling (salt buildup in the condenser tubes) water washing was begun. Wash water containing ammonia was injected into the vapor stream to maintain the pH in the condenser boot between 6.5 and 8. Despite use of an alkaline water wash, a localized region of high H.sub.2 S absorption and H.sub.2 S acidification developed downstream of the water injection nozzle. The resulting iron sulfide corrosion products were a porous scale at this point in the pipe, which allowed more H.sub.2 S to accumulate. The net result was that about 100 square inches of metal was ejected from an 8" pipe elbow about 1.1 m downstream of the water injection nozzle. This depressured the depropanizer into the refinery, forming a vapor cloud which exploded. More details of the failure are reported in the paper entitled "A new form of localized corrosion" in Materials Performance Vol. 32, No 6, June 1993., which is incorporated by reference.
It is also worth noting that although the explosion occurred in May of 1988, it took roughly five years, until June of 1993, to publish the report, which refers to "unpredictable high rates of H.sub.2 S absorption at the turbulent periphery of the water splash zone . . . "
This paper and the disaster exemplify the state of the art in refinery processing, namely controlling water injection based on boot water pH, and relying on inspection to check for localized corrosion problems. Such a retrospective approach can lead to tragic results, though it may be that nothing could have prevented the localized corrosion which occurred in that FCC unit.
We discovered that much of the corrosion that occurs in refinery and chemical process equipment is predictable and avoidable. Before getting into our new method of controlling water washing, it is instructive to review salt formation or introduction and conventional practices associated with salt removal. The review is by no means complete, but illustrates some of the complexities of salt formation and removal. Salt formation in a "wet" process (crude distillation) and a "dry" process (catalytic reforming) will be considered, with some discussion of salt formation mechanisms along the way.
Salt Formation During Crude Distillation--Wet Process
In a typical crude unit, the whole crude passes through one or more stages of heat exchange. The crude is heated to some extent, and heat is recovered from various product and/or reflux streams. The crude is then desalted by mixing with water and separation, which may involve electrostatic precipitation to help remove water droplets. Desalting significantly reduces, but does not eliminate, the salt content of the crude. Desalting can add some undesirable materials to the crude. Desalting uses refinery water streams, which frequently contain dissolved and/or entrained species which can enter the crude oil either by entrainment or via liquid/liquid extraction.
Desalting is never perfect, so significant amounts of salt always remain in the crude. In most crudes these salts are primarily calcium and magnesium chlorides. While desalting involves a water removal stage, there is always some water entrained and/or dissolved in the crude after desalting.
The desalted crude, still containing some calcium and magnesium chlorides and with entrained water added by the desalting step, is then further heated by heat exchange with various hot product streams from the atmospheric tower, the vacuum tower or both. The heat exchanged crude then passes through a direct fired heater to the atmospheric tower where the vaporized distillate is fractionated into various product fractions.
In most refineries stripping steam is added to the atmospheric tower, to product strippers associated with the atmospheric tower, to the vacuum tower and to product strippers associated with the vacuum tower. Steam aids fractionation, in part by creating a "pseudo vacuum". If half the atmosphere in a tower is steam, the hydrocarbon partial pressure is reduced, so that the tower operates as if it were at a lower pressure.
Most refineries recover a gasoline overhead product, naphtha, kerosene, light gas oil, heavy gas oil, and a residual or bottom fraction. The gasoline overhead product is condensed and pumped to a stabilizing system. The liquid streams of naphtha, kerosene etc. pass through strippers and further treating facilities. The bottoms may be used as fuel, or sent to a vacuum distillation unit to recover a vacuum gas oil fraction from a vacuum resid bottoms product.
The crude contains impurities that contribute to ammonium chloride and hydrosulfide salt formation. The impurities generally include sulfur, chlorine and nitrogen compounds which are discussed below.
Sulfur Compounds
Sulfur compounds are present in all crude oils. During high temperature processing, such as occurs in the fired heater and in the lower regions of the crude column, some of the sulfur compounds decompose to form hydrogen sulfide, a key ingredient in hydrosulfide salts.
Chlorine Compounds
Hydrogen chloride can be produced by a variety of reaction mechanisms:
1. hydrolysis of calcium and magnesium chlorides,
2. metathetic reaction between sodium chloride and organic acids, and
3. hydrogenation of organic chlorides.
Nitrogen Compounds
Ammonia enters the crude unit primarily through the desalter wash water. The desalter wash water usually contains about 50 to 100 ppm of ammonia. Since the desalted crude will always have some entrained water, ammonia is invariably carried into the crude tower. Another source of ammonia is hydrogenation of organic nitrogen compounds present in the crude, which can occur to some extent during distillation.
Ammonia and hydrogen chloride can react as they travel up the crude tower to produce ammonium chloride. Copious amounts of ammonium chloride can also be formed in the overhead system itself if ammonia is added as an overhead neutralizer. At lower temperatures this ammonium chloride can precipitate as a solid, directly from the vapor phase, in the upper portions of the atmospheric tower or in the overhead system.
The immediate result is fouling. The fouling can be so severe that there is a significant increase in pressure drop through the column and overhead system. This is usually followed by underdeposit corrosion. Ammonium chloride deposits are hygroscopic, they tend to abstract water from the vapor phase to form a saturated solution of the chloride salt. Saturated solutions of ammonium chloride are acidic and highly corrosive.
Ammonia present in the column, or overhead system, may also react or combine with hydrogen sulfide to form ammonium hydrosulfide, sometimes called ammonium bisulfide. Like ammonium chloride this salt precipitates directly from the vapor phase, forming deposits of ammonium hydrosulfide. Ammonium hydrosulfide is also hygroscopic, so such fouling is usually followed by underdeposit corrosion.
Refiners are aware of the problems of fouling and corrosion in the crude unit and work hard to prevent or at least deal with it. Most refiners use chemical additives such as neutralizer/corrosion inhibitor injection or water wash or a combination of both. The water injection rate is usually based on some "rule of thumb". The typical place for water injection is usually upstream of one of the overhead condensers associated with the column, though usually with no thought to thermodynamic or chemistry considerations.
Plugging of the overhead lines, and less frequently of the tower trays, has been experienced by refiners who picked the wrong injection locations. Additional problems were possible if an incorrect amount of water was added.
If not enough water is injected, or if the flowing stream is too hot and vaporizes all the injected water, then solid salt deposition may occur where there is no aqueous phase present to dissolve the salt. Salts will deposit, followed by corrosion under the salt deposits. Similarly if water is added in the proper amount but too late then fouling may occur upstream of the point of water injection.
Yet another possibility is that enough water will be injected to dissolve salts as they deposit, but not dilute them sufficiently. This produces a corrosive concentrated salt solution. Adding too much water generates unnecessary amounts of waste water.
Unfortunately, the way modern refineries operate there is no way to avoid this problem during crude fractionation. Crude oil always has some salt in it, and some water either dissolved, entrained or in the form of a stable emulsion sometimes called bottoms settlings and water, BS&W. Eliminating desalting might reduce the amount of entrained water attributable to desalting charged to the crude column, but would increase the salt load on the column. Operating with multiple stages of desalting, or hotter water during desalting, would reduce salt content of the crude, but there would still be some salt and likely more water entrainment. There could be more water and more ammonia fed to the crude column even as salt traffic was reduced. This may increase the likelihood of a water phase forming prematurely in the overhead system and promote formation of ammonium compounds.
To summarize, there is no way known to eliminate salts from the crude column. The crude column is inherently wet, at least in the overhead stages, so both salt deposition and salt solution corrosion must be considered.
Such "wet" processes require continuous water washing, but not all processes are wet, nor is continuous water washing always the optimum way to remove salts. If the process with a salt problem is "dry", there usually is no corrosion problem, just a fouling problem. To continuously inject wash water into such a dry process stream creates a corrosion concern and multiplies the volume of waste water generated by the process. The corrosion concern comes about because the aqueous solution formed by water injection can be corrosive. The continuous injection of water can generate very large waste streams.
The process of our invention is directed toward selecting those units where intermittent (as opposed to continuous) water injection is optimum, and also for determining what kind of intermittent water injection program is needed. One type of process which is "dry", and which requires intermittent water injection will be reviewed next, catalytic reforming.
Salt Formation During Reforming--Dry Process
Pt reforming is a dry process. Reformer feed is a clean, hydrotreated material. Even such clean processes can have a significant salt problem, but here much of the salt formation is attributable to the refinery process, catalytic reforming, rather than to the presence of impurities in the feed.
Catalytic reformers pass hydrotreated feed over chlorine containing Pt Catalyst. Chlorine, or possibly some other halogen, is part of the reforming catalyst to impart the desired acidity to the catalyst. Some of this chlorine is "washed" or leached from the catalyst, even though the reforming reaction atmosphere is dry, i.e., has less than 100 ppmv H.sub.2 O. Chlorine reacts with the minor amounts of ammonia, etc. present in reformate to form chlorine salts which deposit in the reformer fractionator impairing its operation.
Refiners have tried to cope with the problems of chlorides in reformate using a variety of approaches, reviewed below.
1. Water Washing
Water washing of a depropanizer associated with a continuous catalytic reformer was reported in Example 2 of U.S. Pat. No. 4,880,568. Periodic water washing for a severe fouling and corrosion problems was not effective so "an elaborate continuous water wash system was installed. The continuous water wash system also failed to solve the deposit problem."
Continuous water washing can create a corrosion problem where none existed before. In a dry atmosphere, chlorine salts are not corrosive, though they will plug equipment. As soon as water is added, a corrosive salt solution forms, and unless all the chlorine salt is removed the salts left unwashed will be soaked with water and highly corrosive.
One of our refineries tried using an aqueous, alkaline treatment of the reformate liquid upstream of the debutanizer. A dilute caustic was injected into reformate intermediate the V/L separator and the debutanizer. The caustic was less than 15.degree. or 20.degree. Be (or roughly 10 to 15 wt % NaOH). A mesh pad was used to aid in separation of caustic/reformate in a separator vessel. The experiment was not considered a success as a flow control valve corroded, and the experiment was stopped.
The engineer responsible for the reformer at this refinery was very concerned about using water injection to remove chlorides, primarily because the dry reformate stream was not corrosive, but it became corrosive if water injection was used. Other approaches considered, solid bed treating and chemical treatment, are reviewed next.
2. Solid Adsorbent Treating
Some refiners use beds of solid adsorbent to prevent chloride corrosion and fouling. More details about this are available from UOP Inc. Des Plaines, Ill., which has approved use of at least one type of solid adsorbent to remove chlorides from reformate.
Solid adsorbent beds can cost a lot. They can also plug, and many refiners are reluctant to use that approach.
3. Chemical Treatments
Several patents are directed at treatment chemicals which can be injected into the reformate stream. These chemicals inhibit the formation of ammonium chloride to keep chlorine compounds in a form which will not precipitate as a solid in process equipment. Some chemical treatment programs also include chelating agents and/or film forming agents to prevent further corrosion.
U.S. Pat. Nos. 5,282,956 and 5,256,276, which are incorporated by reference, disclose inhibiting ammonium chloride deposition by adding an amide such as 1,3-dimethyl-2-thiourea or phosphatide such as lecithin.
U.S. Pat. No. 4,880,568, METHOD AND COMPOSITION FOR THE REMOVAL OF AMMONIUM SALT AND METAL COMPOUND DEPOSITS, Staley et al, Assignee Aqua Process, Inc., Houston, Tex. taught injecting amines and chelating agents into reformate to remove and/or prevent formation of ammonium salt deposits. Amines added form amine salts with a low melting point or an affinity for trace amounts of water. This patent is incorporated by reference.
While adding chemicals to prevent formation of ammonium chloride deposits and/or chelating agents to remove metal corrosion products will help, such approaches are expensive and are not considered the ideal solution. Film forming agents may still be needed to protect metal surfaces in process equipment.
The state of the art could be oversimplified and summarized as follows:
In a dry process, such as reforming, salt removal generally focusses on removing salt deposits before equipment plugs, but without forming a corrosive salt solution in an otherwise dry stream.
Sometimes it is difficult to tell if a process is "wet" or "dry". Most refiners know that Pt catalytic reforming is dry, and that the crude column overhead system is "wet", but it is not easy to determine where other refinery processes fit in this classification scheme.
An incorrect guess can lead to inappropriate treatment. If a unit, or part of a unit is relatively dry, continuous water washing may eliminate a fouling problem but create a corrosion problem (if the salts dissolve in the wash water to form a corrosive solution) and a disposal problem. Thus if a unit is dry enough it is best to let salts deposit and wash the salts from the unit only intermittently. The water washing will still create a corrosive salt solution, but washing may be needed only a few hours a week. The corrosion occurring in such a short period will usually not be significant, and production of waste water from washing can be greatly reduced as compared to continuous water washing.
Although some units are known to be "wet", such as the crude column, and some are known to be "dry", such as reformers, there are many refinery units which are not obviously "wet" or "dry".
We discovered a way to separate refinery and petrochemical units, and pick the ones where an intermittent water injection process would be the optimum way to remove salt and impurity deposits.
Our selection method will also give plant engineers the confidence to pick intermittent water washing, as there is great reluctance in refineries to do something sporadically. Thus many refiners with chloride problems around the reformer, a well known "dry" process, still use continuous water washing, even though we know that this is not necessary and is harmful.
Once intermittent water washing is selected, there are additional concerns that must be addressed. Once water washing starts it must not be stopped until all the deposits are washed out. Leaving some salt deposits in the unit will usually cause under-deposit corrosion, due to the hygroscopic nature of the salts, so complete removal of salt deposits is crucial in an intermittent injection situation.
We have discovered how to solve the problems of intermittent water injection. Thus we calculate:
1. if intermittent, or continuous, water injection is the best way to deal with salt deposition in a particular unit;
2. the amount of water required to wash all the deposits out;
3. a suitable interval, duration and amount of water injection; and
4. the most suitable injection point location.
Thus a way has now been found to select units which will benefit from intermittent water washing, and also to optimize the intermittent water wash procedure.